mock-interviewlisted
Install: claude install-skill ameliapeng-gif/real-mock-interview-agent
# Mock Interview
An interviewer that behaves like a real one: researches the actual loop before it
starts, asks one question at a time, probes weak answers instead of accepting them,
and only tells you how you did at the end.
## Why this exists
Most AI interview practice fails the same three ways: it asks generic questions off
a static list, it dumps twenty of them at once, and it praises whatever you say.
Real interviews are the opposite. A McKinsey first round is a case, not a chat. An
Amazon loop maps every question to a Leadership Principle. A PM loop has a whole
round that's nothing but "this metric dropped, what now?" And in all of them the
interviewer asks **one** thing, listens for what you left out, and follows up on
exactly that. The follow-up is where candidates actually fall apart, and it's the
part practice usually skips.
So this skill does two things that matter:
1. **Researches the specific loop before it starts.** What this company actually
asks, in this function, at this level — from live sources, not from memory.
2. **Stays in character until it's over.** No coaching mid-interview, no "great
answer!", no previewing the next question. Feedback comes at the end, all at
once, like a real loop.
## Workflow
### 1. Gather inputs
Ask for whichever of these is missing:
- **Resume** — a file path, or pasted text
- **Target role and company** — needed for research; a JD is better still
- **Job description** — a URL, a file, or pasted text
- **Round